The next day I went to Ehud and Amir, and I said, “Let’s make Waze the best working place we ever had.” They liked the idea and we defined what it would look like. What mattered for us was that: (1) we support employees and drivers, (2) the founders vote as a single person, and (3) we fire fast if someone doesn’t fit into our culture.

这个就是知识星球对 AI 虽然做了很多探索，但是迟迟不敢放到产品里的原因之一呀。

Project teams had truly mastered “going broad”; they were good at generating lots and lots of ideas . . . But teams did a bad job of picking the best ideas to work on because our selection criteria were generally faulty. The problem stemmed from the team voting on ideas, a classic design thinking approach . . . The ideas with the most votes get explored further. But it turns out that people often vote for what is easy to implement and familiar, and that rarely yields ideas that will surprise and delight customers.

Always bring in early majority users as soon as possible to gather feedback. Recall that they are not going to show up by themselves; you will need to encourage them to try your product.

这也是用户群的重要性。

陈虻说，现在我们有很多纪录片热衷于讲述一个悲欢离合的故事，如果仅仅是这样一个故事，而没有和大的文化背景、时代背景、民族命运相关联的话，其实是背离了纪录片的本源。因为故事片更好看，更能使人动情。现在我们需要解决的一个问题就是：因为走得太远，以至于忘了我们为什么要出发。当我们认真地去研究怎样去拍纪录片的时候，或许已经淡忘了我们为什么要拍纪录片。也就是说，当你过于进入、过于热衷于一个东西时，你就需要放弃这个东西。你只有出去了才能进来，也只有进来了才能出去。现在中国的纪录片恰恰需要跳出去，不要过于陷入，你才能反过来冷静地加以审视；如果一个人过于热爱，这东西就已经不再是它本身，已经变成了你的一种热爱，强加了你许多个人的东西，而不是事件本身。《太极》里说，太想练成的人和三心二意的人，都练不成太极。你必须保持一定状态，才能得到一种真传。按照西方的美学表述就是，距离产生美，必须有一定的距离，贴得太近反而什么也看不见。